Long-term Isopropanol Storage Does Not Alter Fish Morphometrics
Since the discovery of spirit-based preservation, scientists have based innumerable studies of systematics, anatomy, development, biogeography, evolutionary biology, and ecomorphology on fluid specimens held by the world's natural history collections. Though alcohol preservation can alter specimen measurements through dehydration and shrinkage, the magnitude of this problem over decadal timescales has never been estimated. If substantial, long-term preservational effects could compromise studies that depend on the comparability of morphometrics drawn from specimens collected at different times unless the artifacts were explicitly controlled. To evaluate the importance of these potential artifacts, we obtained geometric morphometric data from series of four common fish species collected from the Willamette River drainage over the last 70 years and preserved in the Oregon State Ichthyology Collection in 50% isopropanol after formalin fixation. Most regressions of principal component axes against date of collection revealed no directional trends in specimen morphology, and the two significant regressions lose significance when recent collections from the ecologically distinctive river main stem are excluded. We conclude that specimen shrinkage occurs shortly after preservation and does not exacerbate over time and that much of the observed morphological variation in our sample likely stems from habitat differences among collecting localities. Thus, morphological studies can continue to compare specimens collected over decadal timespans to answer questions about the ecological, evolutionary, and biogeographic causes of morphological variation with minimal concern for major preservational artifacts caused by long-term alcohol storage.
Geographic distribution of examined specimens. Black outline represents area of enlargement on inset political map of the western United States.
(A) Photograph of Richardsonius balteatus showing position of 18 landmarks used in geometric morphometric analysis of lateral body shape. (1) Tip of snout, (2) anterior limit of upper jaw, (3) posterior limit of upper jaw, (4) anterior limit of orbit, (5) ventral limit of orbit, (6) posterior limit of orbit, (7) dorsal limit of orbit, (8) anterior border of epiphyseal bar at midline of dorsal neurocranium, (9) posterior limit of opercle, (10) origin of pectoral fin, (11) origin of dorsal fin, (12) insertion of dorsal fin, (13) dorsal base of caudal fin, (14) posterior margin of caudal peduncle, (15) origin of pelvic fin, (16) origin of anal fin, (17) insertion of anal fin, and (18) ventral base of caudal fin. (B) Photograph of Cottus rhotheus showing position of 14 landmarks used in geometric morphometric analysis of dorsal body shape. (1) Tip of snout, (2) right posterior limit of upper jaw, (3) right opercular spine, (4) insertion of right pectoral fin, (5) origin of first dorsal fin, (6) origin of second dorsal fin, (7) dorsal base of caudal fin, (8) insertion of left pectoral fin, (9) left opercular spine, (10) left posterior limit of upper jaw, (11) anterior limit of right orbit, (12) posterior limit of right orbit, (13) anterior limit of left orbit, and (14) posterior limit of right orbit.
Scatterplot of principal components for each species studied showing substantial overlap in morphospace distributions of specimens grouped by decade. (A) PC1 vs. PC2 for Catostomus macrocheilus, (B) PC2 vs. PC3 for Cottus rhotheus, (C) PC1 vs. PC2 for Rhinichthys osculus, and (D) PC1 vs. PC2 for Richardsonius balteatus.
Deformation grids illustrating shape change on the two most important principal components for each species. First PC for Cottus rhotheus not shown due to its representation of a clear artifact of fixation position. To enhance visualization, left and right panels on each component illustrate twice the observed extreme of variation. Numerical values represent the percent variance explained by each component.
Regressions of shape (PC axes) against year collected for each examined species. Regressions of PC1 show positive, significant linear relationship for Catostomus macrocheilus and Richardsonius balteatus. All other regressions indicate no correlation between shape and time in preservative.
Scatterplots of PC1 versus PC2 for all collection years of (A) Catostomus macrocheilus and (B) Richardsonius balteatus, showing mean morphological difference between specimens from localities in the Willamette River's main channel versus those from tributaries.
Regressions of shape (PC1) against collection year for (A) Catostomus macrocheilus and (B) Richardsonius balteatus after removal of main channel Willamette River specimens, showing no relationship between shape and time in preservative.
Contributor Notes
Associate Editor: W. L. Smith.